Skip to main content

Advertisement

Table 2 Error rate (%) given by linear SVM on MNIST data

From: SDRNF: generating scalable and discriminative random nonlinear features from data

Methods 100 200 300 400 500 1000 1500 2000 2500 3000
RKS 13.64 8.55 6.58 5.96 4.79 3.6 3.48 3.10 2.82 2.27
SDRNF 5.35 3.13 2.23 2.22 2.08 1.62 1.61 1.63 1.59 1.55
  1. The first row indicates different number of approximate features from the kernel matrix. The proposed SDRNF clearly outperforms the other randomized method RKS